Jon Hollinger On Hawks.

[mrhonline]

This seems like a good thread to dump this info (looks around, drops pants, squats) -

I read elsewhere a study of players who overachieved vs. weaker teams/lineups, and Evans was top10. In other words, he's a solid reserve for 18-20 minutes/night, but that's all you should expect of him. He'll likely struggle vs. better defending teams.

He is a journeyman, and there's nothing wrong with that. The one plus in the Hawks' favor is that Evans fills a need that Childress would never have been able to fill - perimeter offense.

Edited August 6, 2008 by ifthenwouldi

[crimedog]

Other factors, such as a lower turnover rate and also Childress does everything in far less minutes than JJ. PER is given to us in per minute fashion, this is why sometimes you see someone with 10 mpg with a PER around 15.

The example I described for player A and B is extremely simplified, PER has a vast number of other factors that contribute to it I just put it in ceteris paribus, so assuming player A and B have every other stat exactly the same we will see player B with a higher PER because of volume shooting over scoring efficiency.

Yeah, but he doesn't have a lower turnover rate. He has a higher turnover rate, quite a bit too...

Its a rate, Joe gets more turnovers but he does A LOT more. In the other thread about GSW and how Al fits in better than Josh because Al is a "low TO player", if you break it down Al actually has a higher percentage of TOs than almost any rotation player on the team... he just doesn't have as much responsibility as a passing hub or scorer though. Its a rate.

Chil vastly outperforms Joe in two areas, scoring efficiency and offensive rebound rate.

[crimedog]

Also, I never said PER does not value efficient scoring, it just rewards a volume shooter more than an efficient scorer. If player A takes only 10 shots inside the arc but makes 60% player A gains 12+2.4-7=7.4. However, if player B takes 25 shots inside the arc but only makes 44% player B gains 22+4.4-17.5=8.9. So while player A is more efficient, player B is the volume shooter who gains the higher PER. Player A is obviously benefitting from being efficient, but player B benefits more because he shoots more.

Its harder to score at volume. That isn't taken into account enough in some metrics (Win score stuff). Kobe shoots in the mid 40s and bynum shot in the low 60s but Kobe is still a superior offensive player. PER has the built in assumption that it is easier to score efficiently at low volumes because you are taking easier shots - which is true, otherwise the only guys that should be taking shots would be Tyson Chandler, Andrew Bynum, and Andris Biedrins.

[Guest]

Yeah, but he doesn't have a lower turnover rate. He has a higher turnover rate, quite a bit too...

I was referring to Chill having a lower turnover rate which helps increase his PER, not JJ.

[crimedog]

I was referring to Chill having a lower turnover rate which helps increase his PER, not JJ.

No, Childress has a higher turnover rate than JJ. 10.8 for Chil and 9.4 for Joe.

Joe scores about twice as much as Chil and gets about 4 times as many assists.

That means that each turnover has less significance...

[Guest]

Its harder to score at volume. That isn't taken into account enough in some metrics (Win score stuff). Kobe shoots in the mid 40s and bynum shot in the low 60s but Kobe is still a superior offensive player. PER has the built in assumption that it is easier to score efficiently at low volumes because you are taking easier shots - which is true, otherwise the only guys that should be taking shots would be Tyson Chandler, Andrew Bynum, and Andris Biedrins.

I have heard that argument before but I don't buy it. Shots are a limited quantity, throughout a game there are roughly 80 field goals throughout a game dependent upon your pace factor. By being a higher volume shooter, you are taking away other opportunities for your teammates. You shouldn't be rewarded for taking away opportunities from your teammates, but that is exactly what PER does in this situation.

[Guest]

No, Childress has a higher turnover rate than JJ. 10.8 for Chil and 9.4 for Joe.

Joe scores about twice as much as Chil and gets about 4 times as many assists.

That means that each turnover has less significance...

You are confusing turnover rate with something that is probably from 82games.

Joe had 223 turnovers in 3343 minutes. That means every 14.991 minutes Joe makes a turnover.

Josh had 98 turnovers in 2274 minutes. That means every 23.2041 minutes Josh makes a turnover.

Josh had a lower turnover rate. Maybe you are looking at turnover % or some sort of hands rating statistic? But for calculating PER, you do not use those statistics. You are given (a value)*turnovers and then at some point in the equation it is divided by minutes played.

[crimedog]

I have heard that argument before but I don't buy it. Shots are a limited quantity, throughout a game there are roughly 80 field goals throughout a game dependent upon your pace factor. By being a higher volume shooter, you are taking away other opportunities for your teammates. You shouldn't be rewarded for taking away opportunities from your teammates, but that is exactly what PER does in this situation.

Giving up on the turnover ratio argument?

Look, if you watch a Hawks game, a GSW game, or a NOH game, you will see 3 guys who score efficiently off of opportunities created by others. Chandler, Biedrins, and Childress get a lot of their points off of offensive rebounds/defensive attention to other players/transition points... (in Chil's case we had this discussion in another thread and ex' pointed out that 56% of Chil's shots come in the first 10 seconds). These are easier shots to make, those guys get credit in PER for being able to get themselves in position for those shots but you can't extrapolate that and say that they should be taking more shots. A huge chunk of Chandler's shots are oops which are hard to miss so he shoots a high percentage, unfortunately another player has to be enough of a threat to draw defensive attention for an alley-oop to work.

Chil & co. get their due in PER for how they score, Joe had a poor first half of the season so his numbers are down quite a bit but the point remains that Chil has a higher PER and we know that it wasn't from his TO ratio (couldn't resist). It is because he was a more efficient scorer. He shouldn't get any more credit for that though because Joe was creating shots for him, not the other way around... Joe was taking harder shots, contested shots, that allows for the high percentage-low volume dudes to do their thing.

I mean, by your logic Josh Childress should be taking the most shots on the team but how would that offense work since so many of his shots are off put backs, transition points, and flashing open.

[foeteen14]

You are confusing turnover rate with something that is probably from 82games.

Joe had 223 turnovers in 3343 minutes. That means every 14.991 minutes Joe makes a turnover.

Josh had 98 turnovers in 2274 minutes. That means every 23.2041 minutes Josh makes a turnover.

Josh had a lower turnover rate. Maybe you are looking at turnover % or some sort of hands rating statistic? But for calculating PER, you do not use those statistics. You are given (a value)*turnovers and then at some point in the equation it is divided by minutes played.

sounds like lawyers playing word games to me. i can see what happens on the court, but then this sneaky lawyer comes in with all his big words and confuses me, now i have reasonable doubt (not Jay Z's first album)

[crimedog]

You are confusing turnover rate with something that is probably from 82games.

Joe had 223 turnovers in 3343 minutes. That means every 14.991 minutes Joe makes a turnover.

Josh had 98 turnovers in 2274 minutes. That means every 23.2041 minutes Josh makes a turnover.

Josh had a lower turnover rate. Maybe you are looking at turnover % or some sort of hands rating statistic? But for calculating PER, you do not use those statistics. You are given (a value)*turnovers and then at some point in the equation it is divided by minutes played.

No, to rate for PER is proportion of possessions ending in a TO.

Look it up on the ESPN website, Hollinger's player profiles have it up, thats where I posted those numbers from. You have to have insider to look.

Here is the quote from Hollinger's stat key:

TO Turnover Ratio is the percentage of a player’s possessions that end in a turnover. Turnover Ratio = (Turnover x 100) divided by [(FGA + (FTA x 0.44) + Assists + Turnovers]

It is what it is, makes more sense. You are wrong, from the horses mouth.

Edited August 6, 2008 by crimedog

[Guest]

I mean, by your logic Josh Childress should be taking the most shots on the team but how would that offense work since so many of his shots are off put backs, transition points, and flashing open.

Whoa whoa whoa, where did I ever said he should be taking the most shots? I have only contested that PER is not a good metric because it overvalues volume shooters. You are taking what I say and creating conclusions that I do not agree with.

You are pointing to specific examples with everything, I am arguing about the formula and not what certain player's PER are. If you want to talk about why PER is so good, say something about what the formula is and not looking at what a certain player's PER is and agreeing with it.

[Guest]

No, to rate for PER is proportion of possessions ending in a TO.

Look it up on the ESPN website, Hollinger's player profiles have it up, thats where I posted those numbers from. You have to have insider to look.

Here is the quote from Hollinger's stat key:

TO Turnover Ratio is the percentage of a player’s possessions that end in a turnover. Turnover Ratio = (Turnover x 100) divided by [(FGA + (FTA x 0.44) + Assists + Turnovers]

It is what it is, makes more sense. You are wrong, from the horses mouth.

Look up the PER formula and point to where Hollinger's "Turnover Ratio" is in the formula. It isn't, its a different statistic.

I explained what turnover ratio I was referring to, so hopefully you can understand it better. Simply put, I should have said turnover rate instead of ratio and that is my mistake. I am not a Hollinger fan so I was unaware that he made up another statistic called that and I see how it could be confusing.

*Quick edit: actually I rescind what I said about confusion with turnover ratio and turnover rate. I looked at what I wrote in my previous post and I clearly said (in response to Exodus) "turnover rate" and not turnover ratio.

Edited August 6, 2008 by hawksfanatic

[crimedog]

Whoa whoa whoa, where did I ever said he should be taking the most shots? I have only contested that PER is not a good metric because it overvalues volume shooters. You are taking what I say and creating conclusions that I do not agree with.

You are pointing to specific examples with everything, I am arguing about the formula and not what certain player's PER are. If you want to talk about why PER is so good, say something about what the formula is and not looking at what a certain player's PER is and agreeing with it.

I'm not saying PER is "so good", I'm drawing a conclusion from your argument and disagreeing with it. You say that you don't think volume should be rewarded. I'm saying that it should because for the most part, the more shots you take/make, the tougher they will be because you will have fewer open shots as you will start to draw more defensive attention.

I don't think its a problem to say that Paul's 20ppg on dribble penetration is more valuable than Chandler's 10ppg on alley-oops even though Chandler's %s are higher.

[crimedog]

Look up the PER formula and point to where Hollinger's "Turnover Ratio" is in the formula. It isn't, its a different statistic.

I explained what turnover ratio I was referring to, so hopefully you can understand it better. Simply put, I should have said turnover rate instead of ratio and that is my mistake. I am not a Hollinger fan so I was unaware that he made up another statistic called that and I see how it could be confusing.

*Quick edit: actually I rescind what I said about confusion with turnover ratio and turnover rate. I looked at what I wrote in my previous post and I clearly said (in response to Exodus) "turnover rate" and not turnover ratio.

There is nothing about "turnover rate" anywhere.

Here is a link to a decent explaination:

http://www.basketball-reference.com/about/per.html

Its based on possessions and normalized to an average number of possessions for the league... its not based on TOs per minute although I suppose that's what I'd calle a "turnover rate".

Edited August 6, 2008 by crimedog

[Guest]

I'm not saying PER is "so good", I'm drawing a conclusion from your argument and disagreeing with it. You say that you don't think volume should be rewarded. I'm saying that it should because for the most part, the more shots you take/make, the tougher they will be because you will have fewer open shots as you will start to draw more defensive attention.

I don't necessarily think that you are wrong that saying the more shots you take the tougher they will be. I don't think that you should be rewarded though. I believe basketball is all about being an efficient player, and from this I value Points per shot over raw points. When you look at PER, it is not making that same connection, all else equal it says the player with more volume shots is rated higher than one who may have more Points per shot.

I find the Win Score to be a better metric. For one, it regresses each statistic over wins and finds linear weights for each statistic. On the other hand, Hollinger uses arbitrary values and also comes into the problem of overvaluing taking an additional shot rather than taking an efficienct shot.

[Guest]

There is nothing about "turnover rate" anywhere.

Here is a link to a decent explaination:

http://www.basketball-reference.com/about/per.html

Its based on possessions and normalized to an average number of possessions for the league... its not based on TOs per minute although I suppose that's what I'd calle a "turnover rate".

When you look at the PER formula, you first see that everything is divided by total minutes played. Then, there are a bunch of other statistics that are added or subtracted together to form the total game score. Later on this is adjusted for team's pace factor and league pace factor and then normalized to 15.

In the link you gave me, after look at the 6th line of the formula which is "- VOP*TOV", that is the only spot in PER that we have turnovers. VOP is just an arbitrary number based off of how the rest of the league performed. This is where you can see that in the formula, it is (1/min)*(assists+points... - X*TO...) where X is positive and arbitrary. So from this you can see that if all else is equal, a player who has more turnovers per minute will have a lower PER than a player who has less turnovers per minute.

And as to "turnover rate", that is just turnovers per minute. Not sure if basketball-reference has a specific statistic for that but that is what I refer to as "turnover rate". Maybe saying turnovers per 48 would have been better?

Edited August 6, 2008 by hawksfanatic

[crimedog]

I don't necessarily think that you are wrong that saying the more shots you take the tougher they will be. I don't think that you should be rewarded though. I believe basketball is all about being an efficient player, and from this I value Points per shot over raw points. When you look at PER, it is not making that same connection, all else equal it says the player with more volume shots is rated higher than one who may have more Points per shot.

I find the Win Score to be a better metric. For one, it regresses each statistic over wins and finds linear weights for each statistic. On the other hand, Hollinger uses arbitrary values and also comes into the problem of overvaluing taking an additional shot rather than taking an efficienct shot.

Well, I'm finally leaving work so this is going to be my last post but the Win Score doesn't pass my "that looks right" test. It uses correlation without ever looking at the results which is why guys like Camby can be in the top 10 above Lebron James...

PER does a good job of pointing out the elite players, Win Score does not.

You still haven't addressed my point that there are only so many easy shots that you can take. Chandler and Biedrins score very efficiently and get rewarded by being rated as above average player by PER. On the other hand, they aren't creating their own shots which is why they can't be higher volume scorers, they just don't have the skillset to be anything other than a finisher. Thats why a guy like Lebron James is leagues ahead of a guy like Andris Biedrins in PER but not in Win Score (I know, you don't like examples but I had to use one just to make the point). In PER, Lebron scores a whole ton of points at a rate above the league average, not above Biedrins' rate but above average. PER takes into account that he must be creating his own shot, and doing an excellent job of it. Win Score doesn't take this into account nearly as much and wonders why Andris Biedrins isn't getting more shots.

[Guest]

Well, I'm finally leaving work so this is going to be my last post but the Win Score doesn't pass my "that looks right" test. It uses correlation without ever looking at the results which is why guys like Camby can be in the top 10 above Lebron James...

PER does a good job of pointing out the elite players, Win Score does not.

You mean that because the outcome agrees with your original thought about who the better players are you agree with it? If that is the case then now I realize why you like PER more than anything else, because they don't agree with your original opinion instead of creating an unbiased metric.

You still haven't addressed my point that there are only so many easy shots that you can take. Chandler and Biedrins score very efficiently and get rewarded by being rated as above average player by PER.

What is there to address? That has nothing to do with the formula, that is a specific example you are referring to. It is pointless to respond to each and every specific example you bring up. Just refer to the formula. You are also extrapolating from this that Win Score or PER "wonders why they don't take more shots". They don't wonder, they are just saying that as a player they are more valuable. You are drawing conclusions that aren't there dude, they don't magically say "Childress should shoot more" they just say he is more valuable because of his efficiency. You are reading what is not there and thinking you are reading between the lines.

[crimedog]

You mean that because the outcome agrees with your original thought about who the better players are you agree with it? If that is the case then now I realize why you like PER more than anything else, because they don't agree with your original opinion instead of creating an unbiased metric.

What is there to address? That has nothing to do with the formula, that is a specific example you are referring to. It is pointless to respond to each and every specific example you bring up. Just refer to the formula. You are also extrapolating from this that Win Score or PER "wonders why they don't take more shots". They don't wonder, they are just saying that as a player they are more valuable. You are drawing conclusions that aren't there dude, they don't magically say "Childress should shoot more" they just say he is more valuable because of his efficiency. You are reading what is not there and thinking you are reading between the lines.

I am of the opinion that that Win Score overvalues certain things because they correlate with winning but it doesn't factor in replacement value. By that I mean better teams have higher FG% but that doesn't mean that the higher efficiency player does more offensively for the team than a high volume player who is of medium efficiency. Not everyone on the team can take uncontested layups or make alley-oops. Win Score rewards those who do and punishes those players who have to score at volume even if they do so in a way that is fairly efficient for the amount of points being scored - thats how Tyson Chandler is ranked above Lebron James in win score. PER takes into account that volume scoring is, to an extent, necessary and harder and gives those players who take the brunt of the offensive responsibility credit. It doesn't say that every high volume scorer is good, Kevin Durant ranked as a barely above average player.

In terms of not liking because I don't agree with it. Its unbiased but that doesn't mean I can't find its results flawed. I just think that Lebron being outside the top 10 raises an eyebrow, Kidd/Camby/Marion being in the top 10 does as well, Okafor and Odom are better than Kobe. It just overvalues rebounding so much. I can watch a game and tell you that Oak and Odom aren't better than Kobe and Kidd, Camby, and Marion are barely playing the same sport as Lebron. Win Score says they are better.

Both metrics are flawed, but I disagree with your reasoning for why PER is wrong and Win Score is better.

You were right about TOs though, good lookin out.

Edited August 6, 2008 by crimedog